No built in safeguard against getting rate limited #2037
Description
Describe the bug
Agents can invoke this MCP server many times in rapid succession and trigger a rate limit lockout from GH api
Affected version
Whatever version Github Copilot Chat uses at the moment at https://github.com/copilot
Steps to reproduce the behavior
- Ask the agent to deep dive into a GH repository
- Observe the agent can trigger a lot of rapid succession invocations of the tool
- Observe a trigger of api rate limiting occurring:
Expected vs actual behavior
If the MCP would trigger a rate limit lockout, it should purposefully slow down and ideally warn the agent to pause between invocations and slow down, or it should just have a built in throttling/backoff to add a cooldown between invocations.
Note
If this is not the github-mcp-server triggering this, feel free to let me know, Im not actually 100% sure this is what is triggering it but it looks like it is, so this seems like the first place to check in.